Glacier clustersGlacier clusters
5.2 Inversion for Fingerprint Magnitudes
5.2.2 Variations of the Ice Sheets
From table5.4it becomes clear that the difference between the total geocentric sea level and relative sea level is in the order of -0.16mm/yr. This is largely explained by the con-tribution of GIA (-0.13±0.04mm/yr). Nerem et al. (2010) applied a larger GIA correction of -0.3mm/yr to obtain relative sea level rise from multi-mission altimetry. Upon read-justing their series with their GIA correction, the trend in the their geocentric sea level is lower (2.02mm/yr) compared to the estimated geocentric sea level trend from this study (2.34mm/yr, see Tab. 5.4). This discrepancy can actually be explained by the contribution of the steric patterns in the Arctic. When the inversion results are propagated to the along track altimetry points, before estimating a uniform sea level rise, a smaller sea level rise is obtained as well (2.16±0.12mm/yr). The latter value agrees to within the error bars with the estimate fromNerem et al.(2010) and with a GSL estimate using altimetry from this study (2.11±0.12mm/yr).
this reason, I choose to set the strength of the regularization relatively low while acknowl-edging that inter-basin (error) correlations exist in the estimates.
Annual Trend W08 ICEsat RMS
A [Gt] ta[doy] [Gt/yr] [Gt/yr] [Gt/yr] post.[Gt]
Green1 <2000m 57.1 (±7.69) 118 (±8) -40.4 (±2.2) -12.0 -14.7 54.2 Green1 >2000m 37.9 (±6.20) 321 (±9) 15.8 (±1.8) -1.0 0.4 43.2 Green2 <2000m 14.8 (±3.39) 115 (±14) -16.8 (±1.0) -6.0 -5.2 23.9 Green2 >2000m 11.5 (±3.50) 47 (±18) 18.3 (±1.0) 19.0 3.0 24.7 Green3 <2000m 37.4 (±5.78) 165 (±9) -16.8 (±1.6) -25.0 -12.2 40.4 Green3 >2000m 27.5 (±5.11) 19 (±11) -1.1 (±1.4) -10.0 2.6 35.9 Green4 <2000m 5.7 (±4.01) 61 (±41) -37.4 (±1.2) -49.0 -43.0 28.7 Green4 >2000m 16.8 (±2.06) 111 (±7) -6.8 (±0.6) -7.0 -0.3 14.8 Green5 <2000m 60.1 (±3.08) 115 (±3) -47.1 (±0.9) -51.0 -55.8 22.7 Green5 >2000m 2.1 (±1.31) 5 (±36) -11.9 (±0.4) 6.0 -4.2 9.1 Green6 <2000m 33.7 (±6.41) 107 (±11) -26.2 (±1.8) -13.0 -24.9 45.6 Green6 >2000m 13.6 (±3.25) 321 (±14) 9.0 (±0.9) 11.0 0.2 22.8 Green7 <2000m 18.1 (±4.58) 131 (±15) -48.7 (±1.3) -14.0 -34.3 32.7 Green7 >2000m 14.0 (±3.84) 364 (±15) 18.1 (±1.1) 2.0 -0.2 26.8 Green8 <2000m 29.0 (±6.33) 93 (±13) -52.5 (±1.8) -16.0 -45.6 47.4 Green8 >2000m 11.8 (±5.73) 244 (±29) -8.3 (±1.7) -13.0 -1.1 41.3 Total >2000m 76.9 (±11.38) 356 (±8) 33.4 (±3.2) 7.0 0.4 78.6 Total <2000m 237.1 (±21.74) 119 (±5) -286.0 (±6.2) -186.0 -235.6 154.4 Total 200.6 (±15.03) 101 (±4) -252.5 (±4.3) -179.0 -235.2 109.0 Table 5.5: Annual mass changes and trends in Gtons for the resolved basins in Greenland.
The trends are furthermore compared to (GRACE) trends fromWouters et al. (2008)(col-umn, W08, using data from Feb 2003 - Jan 2008) and ICEsat derived trends fromSørensen et al.(2011)/Sasgen et al.(2012b), using data from 2003 until 2009.
Generally, we see that at higher elevations (above 2000m) a mass increase, while at the lower elevations a mass decrease is visible. Most obvious are the strong mass losses in the South East of Greenland and in basin 7 where the Jakobshaven glacier is located. At the higher elevations, we see the largest mass increases in the basins 1, 2 and 7. In particular the increase in the high elevation part of basin 1 seems to be a more recent phenomena starting from 2008. The positive mass trend (18.3Gt/yr) of the upper 2000m of basin 2 seems to be a steady phenomena and is also in good agreement with the estimate obtained from ICEsat(19.0Gt/yr).
Apart from the trends, it is clear from Fig. 5.14that several interesting non-secular sig-nals are visible. The low elevation part of basin number 5 in the South East exhibits for example a strong seasonal signal. This is not unexpected, since one expects strong seasonal fluctuations in that region associated with the transport of moist air from the Gulf stream.
Table 5.5 also indicates a strong seasonal signal in the lower section of basin 1, however since the upper section of the basin shows an almost opposite phase, this can be probably be attributed to a correlated error.
The mass loss in many of the basins are accelerating, and some losses appear to be started only recently. When the trends are compared to those of Wouters et al.(2008), which are
derived from a shorter time interval (2003 until the start of 2008), one sees that for example basin 6, 7 and 8 exhibit a significantly larger trend. The acceleration is also confirmed by comparing the trends with those as derived from ICEsat, which uses a somewhat larger time interval (2003-2009). The total ice mass loss in Greenland is estimated to be -252Gt/yr
over the considered time interval (Jan 2003 until Dec 2011).
8
7
6 5 4
3 2 1
−80 −60 −40 −20 0 20 40 60 80
cm/yr
−300−1501503000
Gt
2004 2008 2012
−300−1501503000
Gt
2004 2008 2012
Basin 1
−300
−150 0 150
Gt
2004 2008 2012
−300
−150 0 150
Gt
2004 2008 2012
Basin 2
−150 0 150
Gt
2004 2008 2012
−150 0 150
Gt
2004 2008 2012
Basin 3
−300
−150 0 150
Gt
2004 2008 2012
−300
−150 0 150
Gt
2004 2008 2012
Basin 4
−300
−150 0 150
Gt
2004 2008 2012
−300
−150 0 150
Gt
2004 2008 2012
Basin 5
−300
−1500 150
Gt
2004 2008 2012
−300
−1500 150
Gt
2004 2008 2012
: <2000m : >2000m
Basin 6
−450−300
−1501500
Gt
2004 2008 2012
−450−300
−1501500
Gt
2004 2008 2012
Basin 7
−450−300
−1501500
Gt
2004 2008 2012
−450−300
−1501500
Gt
2004 2008 2012
Basin 8
Figure 5.14: Estimated basin changes in Greenland in Gt. The background image depicts the computed trend in terms of equivalent water height. The trend is obtained by sum-ming all basin contributions in terms of geoid height. Subsequently, a mean, trend and annual harmonic is fitted through each spherical harmonic coefficient. The obtained trend is then converted from geoid height to surface load and evaluated in the spatial domain.
Antarctica and GIA
As stated earlier in Sec. 4.2.4, the challenge in Antarctica is to separate the GIA signal from the present day mass changes in the drainage basins. It is therefore useful to discuss these topics together.
Similar to the discussion of the Greenland mass changes, a table (5.6) is provided where the annual fits and trends of the basins are assembled. To obtain an impression of the time
variations, the estimated basin curves have been plotted over time in Fig. 5.15. The esti-mated trend is plotted in the spatial domain in the same figure.
From Fig.5.15, it is obvious that the most prominent changes occur in the Amundsen sea sector (basins 20-23). These basins exhibit mass losses from 30 to over 60Gt/yr. It should be stressed that these changes are mainly related to the accelerating glacier velocities and melt-ing in this region, and are virtually unaffected by the GIA signal, which only contributes an apparent 1-2Gt/yrper basin. Furthermore, although a significant annual signal is present, the long term signal is dominated by trends and more recent accelerations. Compared to the results of Sasgen et al. (2012a) the estimated trends from this study are consistently larger by about 10-30Gt/yr. although the time period is similar. In the current study, the GRACE signal up to degree and order 150 is used, and only a weak inter-basin constraint is applied. This possibly allows more signal to propagate in the estimates, compared to using filtered GRACE RL04 data.
5 5 5
7
−60 −40 −20 0 20 40 60
cm/yr
1 2
3 4
5 6
7 8 9 10
11 12
13
14 15
16 18 17
19 20 21 22 23
24 25
26 27
−150 0 150
2004 2008 2012
−150 0 150
2004 2008 2012
24 25 26 27
−150 0 150 300
2004 2008 2012
−150 0 150 300
2004 2008 2012
1 18 19
−600
−450−300
−1500 150
2004 2008 2012
−600
−450−300
−1500 150
2004 2008 2012
20 21 22 23
−150 0 150
2004 2008 2012
−150 0 150
2004 2008 2012
14 15 16 17
−150 0 150 2004 2008 2012
−150 0 150 2004 2008 2012
10 11 12 13
−150 0 150 2004 2008 2012
−150 0 150 2004 2008 2012
6 7 8 9
−150 0 150
2004 2008 2012
−150 0 150
2004 2008 2012
2 3 4 5
Figure 5.15: As in Fig. 5.14, but now for the Antarctic ice sheet. The blue contours indicate the estimated GIA uplift (see also Fig.5.16).
The estimated GIA signal, in terms of present day uplift, is displayed in Fig. 5.16. The plot has been constructed by simply adjusting the spherical harmonic coefficients of the
Annual Trend S12 GIA RMS
A (Gt) ta(doy) Gt/yr Gt/yr Gt/yr post.
02_EAIS 13.1 (±1.79) 286 (±8) -11.5 (±0.5) -6 9.0 12.9 03_EAIS 14.8 (±2.92) 261 (±12) 1.8 (±0.9) 4 8.0 21.1 04_EAIS 23.3 (±3.08) 203 (±7) 13.2 (±0.9) 12 1.0 21.5 05_EAIS 20.2 (±3.69) 11 (±10) 8.0 (±1.0) 5 0.0 27.2 06_EAIS 16.8 (±2.59) 235 (±9) 4.7 (±0.8) 5 2.0 18.8 07_EAIS 6.4 (±5.04) 336 (±45) 21.7 (±1.4) 13 2.0 35.0 08_EAIS 5.5 (±3.48) 360 (±35) 12.9 (±1.0) 15 1.0 25.2 09_EAIS 5.9 (±2.82) 180 (±27) -7.0 (±0.8) -1 1.0 19.9 10_EAIS 4.0 (±1.88) 292 (±28) -2.7 (±0.6) -1 5.0 13.9 11_EAIS 2.0 (±2.48) 250 (±73) -0.1 (±0.7) 9 1.0 17.7 12_EAIS 19.5 (±3.62) 275 (±11) 4.1 (±1.1) -8 4.0 26.1 13_EAIS 14.4 (±4.21) 251 (±17) -9.7 (±1.2) -8 5.0 30.4 14_EAIS 11.0 (±4.72) 215 (±25) -14.3 (±1.4) -8 2.0 31.8 15_EAIS 5.7 (±2.48) 96 (±26) -6.5 (±0.7) -2 0.0 17.8 16_EAIS 6.0 (±1.83) 355 (±17) 3.7 (±0.5) -5 1.0 12.4 17_EAIS 19.1 (±3.84) 253 (±12) -18.8 (±1.1) -4 13.0 27.6 01_WAIS 6.1 (±4.13) 68 (±40) 33.2 (±1.2) 9 6.0 29.8 18_WAIS 7.9 (±2.17) 272 (±17) -5.0 (±0.6) 8 6.0 15.6 19_WAIS 10.5 (±2.64) 235 (±15) 31.7 (±0.8) 8 6.0 18.9 20_WAIS 5.6 (±7.37) 185 (±74) -56.1 (±2.1) -38 1.0 49.7 21_WAIS 8.3 (±4.03) 311 (±28) -62.4 (±1.2) -51 2.0 29.5 22_WAIS 9.8 (±5.18) 253 (±32) -31.1 (±1.5) -25 2.0 36.4 23_WAIS 5.0 (±4.51) 175 (±50) -41.4 (±1.3) -12 0.0 31.4 24_PENIN 10.8 (±3.77) 263 (±21) 13.9 (±1.1) 4 1.0 26.6 25_PENIN 5.3 (±3.21) 72 (±36) -12.4 (±0.9) -25 0.0 22.8 26_PENIN 14.0 (±3.13) 27 (±13) -12.0 (±0.9) - 0.0 21.9 27_PENIN 9.3 (±3.94) 102 (±25) -4.5 (±1.1) - 1.0 27.4 PENIN 14.4 (±7.55) 43 (±30) -15.1 (±2.1) -21 2.0 53.4 WAIS 29.1 (±16.86) 247 (±34) -131.2 (±4.9) -102 22.0 119.4 EAIS 107.6 (±20.28) 262 (±11) -0.5 (±6.0) 19 57.0 146.9 Total 123.0 (±33.64) 263 (±16) -147.0 (±9.9) -103 81.0 237.9
Table 5.6: Annual mass changes and trends in Gtons for the drainage basins in Antarctica.
The column denoted with S12 denote the values as published bySasgen et al.(2012a), who used the same basin delineation. The apparent mass change as would have been induced by the estimated GIA model is tabulated in the column marked with ’GIA’.
These values are computed by means of basin averaging the estimated GIA signal.
reference GIA uplift using the estimated GIA parameters. In the spectral domain, such a rescaling is justified since the relationship between the GIA-induced Stokes coefficients and the associated uplift coefficients is approximately linear (Wahr et al.,2000,Purcell et al., 2011).
Estimated GIA Uplift (U)
Estimated GIA Uplift (U) Estimated GIA Uplift (U)Estimated GIA Uplift (U)
−10
−8
−6
−4
−2 0 2 4 6 8 mm/yr10
Estimated update ∆U
Estimated update ∆U Estimated update Estimated update ∆∆UU
−4
−3
−2
−1 0 1 2 3 mm/yr4
Figure 5.16: Estimated GIA uplift in the northern and southern hemisphere. The bottom two figures, indicate the change relative to the a priori model (ICE-5G, with VM2 earth model).
Compared to the reference GIA model in Antarctica, the estimated model shows 31%
smaller amplitudes. Smaller GIA signals were also found byWhitehouse et al.(2012) and Sasgen et al.(2012a). In particular,Sasgen et al. (2012a) suggested that the GIA signal is much smaller than previously assumed. Integrated over Antarctica they found an
appar-ent mass change of 48Gt/yr which is roughly twice as small as the value estimated here 81Gt/yr. Accordingly, this difference is also reflected in the total present day mass change.
Interesting is that, when the GIA constraints , applied in this study, would be loosened such values would also be obtained. It is therefore likely that the GIA constraint over Antarctica might still be too strong, but a more detailed study is not within the scope of this work.
One thing that would speak against this, is that the basins 1 and 19, which are now asso-ciated with a strong positive trend (roughly 30Gt/yr), would be even stronger when the GIA component was decreased. Alternatively, it could be that a different GIA pattern, whose maximum values are shifted towards the centers of basin 1 and 19, would fit the data better.
The basins which are most affected by the GIA signal lie on the boundary between West and East Antarctica (basins 1, 2, 3, 17, 18, 19). Although the GIA signal in basins 2, 3 and 17 are not strong in value, they may still falsely contribute to the mass imbalance of Antarctica, since the basins are so large. It should be remarked that the ice loading history of ICE5-G results in significantly more signal in east Antarctica compared to the ice loading histories fromIvins and James(2005),Whitehouse et al.(2012),Shepherd et al.(2012).
Compared to the trends, inter-annual variations play a more important role in East Antarc-tica and on the Antarctic peninsula. These are likely associated with atmospheric events.
For example the recent changes in Dronning Maud land (basins 4-7) have been associated with large scale snow fall events (Böning et al.,2012).